English:Ethics of Artificial Intelligence
Ethics of Artificial Intelligence
Introduction
Artificial intelligence is already part of everyday life. Recommendation systems suggest videos, navigation apps find routes, cameras can detect faces, translation tools convert languages, and generative AI can create text, pictures, sound, and code. AI can be useful, but useful technology can also create problems when it affects people unfairly, collects too much data, spreads false information, or makes important decisions without enough human control.
AI ethics is the study of how AI should be designed, developed, and used so that it supports people and reduces avoidable harm. It asks questions such as: Who benefits? Who could be harmed? Is the system fair? What personal data does it use? Can people understand what it is doing? Who is responsible when something goes wrong?
For Grades 7–8, the most important goal is not to memorize a list of rules. It is to learn how to notice an ethical problem, identify the people affected, compare possible benefits and harms, ask for evidence, and explain a responsible choice.
UNESCO's Recommendation on the Ethics of Artificial Intelligence places human rights and human dignity at the center of AI ethics and highlights principles such as fairness, transparency, environmental responsibility, and human oversight. UNICEF's guidance on AI and children emphasizes safety, privacy, fairness, transparency, accountability, inclusion, and children's best interests. These ideas will guide this aiMOOC.
What Makes an AI Question Ethical?
An AI question becomes an ethical question when it concerns what people should do, what is fair, whose rights matter, or how benefits and harms should be shared. Technical questions and ethical questions often connect.
For example, a technical question might ask, "How accurately can this system identify a face?" An ethical question asks, "Should facial recognition be used here at all, and what happens if it is wrong?" A system can work as designed and still be used in a way that is unfair or unsafe.
You can use four simple steps when you meet an AI ethics problem:
- Identify the people affected: Who uses the system, who is judged by it, and who may never have agreed to be part of it?
- Compare benefits and harms: What could improve, and what could go wrong?
- Check the evidence: What data, tests, or reliable sources support the claims?
- Choose responsibility: Who should decide, monitor, explain, and correct the system?
Ethics does not always give one easy answer. Two values can conflict. A school may want an AI security system to make a building safer, while students may also value privacy. Good ethical reasoning explains the trade-off instead of pretending that only one value matters.
AI Is Not a Moral Shortcut
AI systems do not remove human responsibility. People choose goals, collect or select data, build models, decide where systems are used, and create rules for their use. Even when an AI system produces an output automatically, people and organizations still have responsibilities for how that output is used.
A useful habit is to ask: Who made the important choices before the AI produced this result? That question often reveals where responsibility belongs.
Core Principles of Responsible AI
Fairness and Bias
Fairness means that people should not receive unjustly different treatment because of irrelevant characteristics or poorly designed rules. In AI, unfair results can come from many places: unrepresentative training data, biased historical records, labels created by people, missing information, badly chosen goals, or a system used in a situation for which it was not designed.
Bias does not always mean that a programmer deliberately tried to discriminate. A dataset can reflect patterns from an unequal society. A sample can leave out important groups. A measurement can work better for some people than for others. This is why testing an AI system on different groups and situations matters.
The image above uses art to draw attention to algorithmic bias. Ethical thinking often needs both data and imagination: numbers can reveal a pattern, while stories and art can help you understand how that pattern affects real people.
Facial Recognition as a Fairness Example
Face detection and facial recognition are not the same thing. Face detection locates a face in an image. Facial recognition tries to identify or verify a person. Both can raise ethical questions when used in public spaces, schools, policing, or security.
Suppose a facial recognition system works less accurately for one group of people. If the system controls access to a service, those errors may become unfair barriers. If it is used for a serious decision, a false match can cause much greater harm. The ethical question is therefore not only "What is the average accuracy?" but also "For whom does it fail, how serious are the errors, and what human review exists?"
Joy Buolamwini's work on facial analysis helped make algorithmic bias visible to a wide public audience.
Privacy and Data Protection
AI systems often depend on data. Some data is harmless, while other data can reveal sensitive details about a person's identity, location, habits, health, family, or beliefs. Privacy is about having appropriate control over personal information and being protected from unnecessary observation or misuse.
Before sharing data with an AI tool, ask:
- What data is being collected: Is your name, voice, face, location, schoolwork, or device information included?
- Why is it needed: Is the data necessary for the task?
- How long is it kept: Can it be deleted later?
- Who receives it: Is it sent to another company or service?
- What choice do you have: Can you understand and refuse the collection when appropriate?
For students, a strong rule is simple: do not enter private information about yourself or other people into an AI service unless you are allowed to do so and understand how the data will be used.
Transparency and Explainability
Transparency means being open about the fact that AI is being used and providing useful information about its purpose, limits, data, and decision process. Explainability means helping people understand why an AI system produced a result.
Not every advanced AI system can give a complete step-by-step explanation of its internal calculations. However, users can still be told important things: what the system is for, what information it uses, how well it was tested, what its limits are, and how a person can question or appeal a decision.
A classroom example is an AI writing assistant. Students should know that the tool can make mistakes and may invent plausible-sounding information. Teachers should also be clear about when AI help is allowed and how students should show their own work.
Accountability and Human Oversight
Accountability means that people or organizations can be held responsible for decisions and outcomes. Human oversight means that people can monitor, question, stop, or correct an AI system when necessary.
High-impact decisions need especially careful oversight. If an AI system recommends who receives a scholarship, who gets medical attention first, or whether a vehicle should brake, people need clear rules about who checks the system and who acts when it fails.
A self-driving vehicle shows why responsibility can be complicated. Engineers build the system, a company operates it, regulators set rules, and a human passenger may have limited control. Ethical analysis asks how safety is tested, who is protected, how failures are reported, and who is accountable for decisions.
Safety, Reliability, and Security
A responsible AI system should be tested for the situation in which it will be used. Reliability means that it performs consistently enough for its purpose. Safety means reducing the chance of physical, psychological, social, or financial harm. Security means protecting the system from attacks, misuse, or unauthorized access.
An AI tool that is safe for suggesting music is not automatically safe for giving medical advice. The level of testing and human review should match the possible harm.
NIST describes trustworthy AI using characteristics that include validity and reliability, safety, security and resilience, accountability and transparency, explainability, privacy, and fairness with harmful bias managed.
Human Rights, Inclusion, and Accessibility
Ethical AI should respect human rights and should not treat people as less important because of disability, gender, race, language, income, religion, age, or another protected or personal characteristic. Inclusion means involving different communities in design and testing, especially people who may be strongly affected by a system.
Accessibility also matters. An AI service that only works well for one language, one type of device, or one way of seeing and hearing may exclude people. Designing for a wider range of users can improve both fairness and usefulness.
Generative AI, Deepfakes, and Information Integrity
Generative AI can create new text, images, audio, video, and code. These tools can support creativity and learning, but they can also produce false claims, copied-looking work, misleading images, or realistic media that makes a person appear to say or do something they never did.
The timeline above shows how quickly AI-generated faces became more realistic. A realistic image is not proof that an event happened or that a person exists.
The person in the image above is synthetic. This is a useful reminder that appearance alone is not enough to judge whether media is authentic.
A deepfake is synthetic or altered media created with AI techniques so that a person appears to say or do something they did not actually say or do. Deepfakes can be used for entertainment or creative work, but they can also be used for bullying, impersonation, scams, or misinformation.
When you see surprising AI-related content, use media literacy:
- Check the source: Who published it, and is that source reliable?
- Look for independent evidence: Do trustworthy sources report the same event?
- Check date and context: Is an old image being presented as new?
- Investigate the media: Search for earlier versions when possible.
- Delay sharing: Do not spread content just because it is dramatic.
AI in School
AI can explain a concept, generate practice questions, translate text, brainstorm ideas, or help with accessibility. It can also create problems if it replaces learning, hides who did the work, exposes personal data, or produces wrong information.
A responsible student can use this rule: AI may support your thinking, but it should not secretly replace your thinking.
Before using AI for schoolwork, ask:
- Is this use allowed: Follow your teacher's and school's rules.
- Can I verify the result: Check important claims with reliable sources.
- Can I explain the work myself: You should understand what you submit.
- Am I protecting personal data: Do not upload private information without permission.
- Should I disclose AI help: Follow the required citation or disclosure rules.
UNESCO notes that AI in education can create opportunities while also creating risks, so ethical use should protect learners' rights and support human-centered education.
Environmental and Social Effects
AI ethics also includes effects that are easy to miss. Large digital systems use electricity, computing hardware, data centers, and water for cooling in some settings. Hardware production also uses materials and creates electronic waste. The exact environmental impact depends on the system, energy source, hardware, location, and how often the system is used.
Social effects matter too. AI can change jobs, influence what information people see, shape access to services, and alter how creative work is made. An ethical evaluation should therefore ask not only "Can we build this?" but also "Who gains, who pays the costs, and what alternatives exist?"
A Simple AI Ethics Check
When you evaluate an AI system, use the word FAIRER as a memory aid:
- Fairness: Does it treat people justly, and has bias been tested?
- Accountability: Who is responsible for the system and its outcomes?
- Information: What data and evidence does it use?
- Rights: Does it respect privacy, dignity, access, and other rights?
- Explainability: Can users understand the purpose, limits, and important reasons?
- Review: Is there meaningful human oversight and a way to correct mistakes?
This checklist is not a law and does not solve every dilemma. It is a practical way to organize your questions before making a judgment.
Ethical Case Studies
Case Study: The Homework Detector
A school buys an AI system that estimates whether homework was written with generative AI. The company says the detector is useful but cannot guarantee perfect accuracy. A student is accused of cheating because the detector gives a high score.
Ethical questions include: Should one automated score be enough to punish a student? What evidence should the student be allowed to see? How can the student appeal? Who is responsible for checking false positives? A fair process would use multiple forms of evidence and meaningful human review rather than treating an uncertain score as proof.
Case Study: The Smart Camera
A school wants cameras that automatically identify people entering the building. The goal is safety, but the system would process facial data from students, staff, families, and visitors.
The ethical trade-off includes safety, privacy, consent, accuracy, data retention, security, and possible unequal error rates. A responsible decision would compare the technology with less intrusive alternatives and ask whether the expected benefit is strong enough to justify the privacy cost.
Case Study: The AI Tutor
An AI tutor gives fast explanations and adapts exercises to each student. It may help students who need extra practice. However, it may also produce incorrect explanations, collect detailed learning data, or encourage a student to depend on it too much.
A responsible classroom plan would keep teachers involved, protect student data, check the quality of explanations, and teach students to question the system rather than trusting it automatically.
Global Guidance
AI ethics is discussed around the world. Different organizations use different terms, but several ideas appear repeatedly: respect for human rights, fairness, privacy, transparency, safety, accountability, and human oversight.
UNESCO's 2021 Recommendation on the Ethics of Artificial Intelligence is a global framework centered on human rights and dignity. UNICEF updated its Guidance on AI and Children in 2025 and lists child-centered requirements including safety, privacy, non-discrimination, transparency, accountability, inclusion, and children's well-being. The OECD AI Principles emphasize human rights, democratic values, fairness, privacy, transparency, robustness, safety, and accountability. NIST's AI Risk Management Framework helps organizations think systematically about trustworthy AI and risk.
These frameworks are useful because ethical AI is not only about personal behavior. Schools, governments, companies, researchers, designers, and communities all have responsibilities.
Reliable Sources for Further Study
- UNESCO Recommendation on the Ethics of Artificial Intelligence: A global framework centered on human rights and dignity.
- UNICEF Guidance on AI and Children: Child-centered guidance on safety, privacy, fairness, transparency, accountability, inclusion, and well-being.
- OECD AI Principles: International principles for trustworthy and human-centered AI.
- NIST AI Risk Management Framework: A framework for identifying and managing AI risks and trustworthiness.
Interactive Tasks
Quiz: Test Your Knowledge
What is the main purpose of AI ethics? (To guide responsible choices about how AI is designed and used) (!To make every AI system fully automatic) (!To remove all human decision making) (!To make computers think exactly like humans)
Which situation can create unfair AI results? (Training data that leaves out important groups) (!A computer using electricity) (!A screen showing large text) (!A keyboard having many keys)
What does privacy mainly protect? (Personal information and control over its use) (!The speed of a computer processor) (!The color of a website) (!The size of a monitor)
What is transparency in responsible AI? (Being open about when AI is used and what it is for) (!Keeping every AI decision secret) (!Making AI systems impossible to question) (!Removing all information for users)
What does accountability mean in AI ethics? (People or organizations are responsible for outcomes) (!The AI system is always responsible by itself) (!No one needs to correct mistakes) (!Only users can be responsible)
Why is human oversight important? (People can review and correct important AI decisions) (!It guarantees that AI never makes errors) (!It lets AI ignore safety rules) (!It removes the need for evidence)
What is a good response to a surprising deepfake video? (Check the source and look for independent evidence) (!Share it immediately with everyone) (!Assume realistic video is always true) (!Trust it because it has many views)
What is a responsible way to use AI for schoolwork? (Use it within school rules and verify important claims) (!Submit every AI answer without reading it) (!Upload private information about classmates) (!Hide all AI help from the teacher)
Why does a self driving vehicle raise ethical questions? (Its decisions can affect safety and responsibility) (!It always travels faster than every driver) (!It never needs testing) (!It has no effect on people)
Which idea is part of child centered AI guidance? (Protect children from unfair treatment and privacy harms) (!Collect as much child data as possible) (!Hide when children are interacting with AI) (!Remove adults from important safety decisions)
Memory Game
| Fairness | Treating people without unjust discrimination |
| Privacy | Protecting personal information and control over it |
| Transparency | Making it clear when and how AI is being used |
| Accountability | Having people or organizations responsible for outcomes |
| Bias | A systematic tendency that can create unfair results |
| Oversight | Human checking and control of important AI decisions |
| Consent | Permission given after understanding what will happen |
| Deepfake | AI generated or altered media that can imitate real people |
Drag and Drop
| Match the correct terms. | Topic |
|---|---|
| Training data | Examples used to teach an AI system patterns |
| Human oversight | People review important AI decisions |
| Explainability | Reasons for an AI output can be understood |
| Data minimization | Collect only the personal data that is needed |
| Fact checking | Verify a claim with reliable independent evidence |
...
Crossword Puzzle
| Fairness | What principle asks whether people are treated justly |
| Privacy | What principle protects personal information |
| Bias | What can cause systematic unfair results |
| Transparency | What principle asks for openness about AI use |
| Accountability | What principle asks who is responsible for outcomes |
| Oversight | What word means human monitoring and control |
LearningApps
Cloze Text
Open-Ended Tasks
Easy
- AI in My Day: Keep a one-day diary of places where you notice AI or automated recommendations, then choose one example and explain one benefit and one possible ethical concern.
- Privacy Poster: Create a clear poster for students your age showing five questions to ask before sharing personal information with an AI tool.
- Fairness Scenario: Write a short classroom story in which an automated rule treats someone unfairly, then rewrite the rule to make the decision process fairer.
- Fact Check Challenge: Ask an AI system for a factual explanation of a school topic, verify three claims with reliable sources, and mark which parts were correct, unclear, or wrong.
Standard
- AI Interview: Interview a teacher, parent, librarian, or other adult about one AI system they use, then summarize what data it uses, who benefits, and what concerns they have.
- Technology Walk: With teacher permission, visit your school library, office, or IT area and identify one digital system that handles personal data, then ask what privacy safeguards are used.
- Deepfake Awareness Video: Create a sixty to ninety second video that teaches classmates how to pause, check the source, and look for independent evidence before sharing suspicious media.
- School AI Code: Work with a small group to draft six classroom rules for responsible AI use covering learning, privacy, fairness, verification, disclosure, and human responsibility.
Advanced
- Stakeholder Debate: Research a proposed use of facial recognition in a school and prepare arguments from the perspectives of students, families, teachers, school leaders, and privacy advocates before proposing a balanced policy.
- Autonomous Vehicle Ethics Board: Analyze a realistic self-driving vehicle safety scenario, identify the stakeholders and uncertainties, and write a recommendation explaining how safety testing and accountability should work.
- Bias Audit Experiment: Build a small fictional dataset for a school club selection system, test how changing missing or unbalanced examples affects the outcome, and explain why the experiment does not prove how a real AI system would behave.
- Responsible AI Project: Design a prototype idea for an AI tool that could help your school, then create a risk-benefit report covering fairness, privacy, safety, accessibility, environmental cost, and human oversight.
Learning Assessment
- Case Analysis: Given a new AI use in school, identify at least four stakeholders, explain two benefits and two risks, and justify whether the system should be used, changed, tested further, or rejected.
- Fairness Evaluation: Compare two fictional AI systems with different error patterns and explain why average accuracy alone may not be enough to judge fairness.
- Privacy Decision: Examine a data-collection plan for an AI learning app and recommend which data should be collected, avoided, protected, or deleted, with reasons.
- Evidence Check: Evaluate an AI-generated claim by comparing it with at least two independent reliable sources and explain which evidence changed your judgment.
- Responsibility Map: For an autonomous or automated system, map the responsibilities of developers, operators, users, and decision makers when a serious error occurs.
- Transfer Task: Apply the FAIRER ethics check to an AI system you have not discussed in class and explain which principle creates the hardest trade-off.
Evidence of Learning
- Knowledge: You can explain fairness, bias, privacy, transparency, explainability, accountability, safety, and human oversight in your own words.
- Reasoning: You can identify stakeholders, compare benefits and harms, notice uncertainty, and explain ethical trade-offs instead of giving unsupported opinions.
- Source Skills: You can verify AI-related claims with reliable independent evidence and recognize that realistic media can still be synthetic.
- Products: You can create a poster, report, policy, video, debate brief, or data experiment that communicates responsible AI ideas clearly.
- Collaboration: You can listen to different stakeholder perspectives, disagree respectfully, and revise a proposal when stronger evidence appears.
- Transfer: You can apply ethical principles to a new AI system in school, transport, media, health, public services, or another unfamiliar context.
OERs on the Topic
Linked Learning Areas